- Email: [email protected]
Leukocyte ABCA1 gene expression is associated with fasting glucose concentration in normoglycemic men
Leukocyte ABCA1 Gene Expression Is Associated With Fasting Glucose Concentration in Normoglycemic Men C. Albrecht, I. Simon-Vermot, J.I. Elliott, C.F. Higgins, D.G. Johnston, and J. Valabhji Adenosine triphosphate (ATP)-binding cassette transporter A1 (ABCA1) mediates the efflux of cholesterol to apolipoprotein A1, a process necessary for high-density lipoprotein (HDL) formation and reverse cholesterol transport. In patients with Tangier disease, mutations in ABCA1 result in low circulating HDL-cholesterol and predisposition to coronary heart disease (CHD). ABCA1 gene expression is decreased in diabetic mice. In humans, glycated hemoglobin (HbA1c) predicted future CHD events, even within the normal range. We hypothesised that leukocyte ABCA1 gene expression would be inversely associated with indices of glycemia in normoglycemic men. Fasting blood samples were taken from 32 healthy, nonsmoking, normoglycemic men (age 23 to 46 years). ABCA1, peroxisome proliferator–activated receptor gamma (PPAR␥), and liver X receptor alpha (LXR␣) gene expressions in circulating leukocytes were measured using TaqMan technology. Significant inverse associations between ABCA1 gene expression and both fasting glucose concentration (r ⴝ ⴚ0.49, P ⴝ .008) and age (r ⴝ ⴚ0.39, P ⴝ .043) were found. There was no association with HbA1c (r ⴝ ⴚ0.23, P ⴝ .238) or HDL-cholesterol concentration (r ⴝ 0.02, P ⴝ .904). In a multiple regression model, fasting glucose remained a significant independent predictor (P ⴝ .037), whereas age did not (P ⴝ .226). Mechanisms underlying the association were explored; there were no significant associations between fasting glucose concentration and leukocyte PPAR␥ gene expression, or between fasting glucose concentration and leukocyte LXR␣ gene expression. This is the first demonstration of an association between ABCA1 gene expression and fasting glucose concentration in vivo. © 2004 Elsevier Inc. All rights reserved.
A
LOW CONCENTRATION of high-density lipoprotein (HDL)-cholesterol is a powerful and independent predictor of coronary heart disease (CHD),1,2 and this is thought, though not proven, to relate to the role of HDL in reverse cholesterol transport. Recent studies have demonstrated that adenosine triphosphate (ATP)-binding cassette transporter A1 (ABCA1) is highly regulated in macrophages and directly or indirectly mediates the efflux of cholesterol and phospholipids to apolipoprotein A1, a process necessary for HDL formation, and the first step in reverse cholesterol transport.3-5 Whether ABCA1 is itself the cholesterol transporter or whether it regulates the transport process is not yet known. The identification of mutations in ABCA1 in patients with Tangier disease and familial HDL deficiency6-8 demonstrated that inadequate transport of phospholipid and cholesterol to the extracellular space results in the hypercatabolism of lipid-poor nascent HDL particles, low circulating HDL-cholesterol, and the predisposition to CHD. More common variants in the gene encoding ABCA1, such as the R219K variant, which has a carrier frequency of 46% in Europeans, are also associated with altered lipid levels and a modified risk of CHD.9 The association between ABCA1 and CHD is only partly explained by HDL-cholesterol concentrations. Indeed, in apparently normal human populations, an association between ABCA1 and HDL-cholesterol has not been clearly demonstrated. Common variants in the gene encoding ABCA1 are not associated with changes in HDL-cholesterol, despite modifying CHD risk.10 ABCA1 regulatory variants influence CHD-independent of effects on HDL-cholesterol.11 In low-density lipoprotein (LDL)-receptor knockout mice, leukocyte ABCA1 plays a critical role in the protection against atherosclerosis, but disruption of leukocyte ABCA1 function does not affect plasma HDL-cholesterol concentrations despite increasing the atherosclerotic risk.12 Experiments involving bone marrow transplantation in ABCA1-deficient and wild-type mice suggest that macrophage expression of ABCA1 makes minimal contribution to plasma HDL concentrations.13 Diabetes is associated with increased CHD risk.14,15 The risk Metabolism, Vol 53, No 1 (January), 2004: pp 17-21
is in excess of that predicted by lipid parameters, blood pressure, smoking status, and age. Whether hyperglycemia itself mediates the excess risk is unproven. ABCA1 gene expression is decreased in the liver and peritoneal macrophages of diabetic compared to control mice.16 Reduced ABCA1 function could therefore contribute to the excess CHD risk in human diabetes, and this could be independent of HDL-cholesterol. The potential importance of glucose itself is suggested by a recent prospective population-based study in men, the Norfolk cohort of the European Prospective Investigation into Cancer and Nutrition. This study has demonstrated that glycated hemoglobin (HbA1C), a measure of average circulating glucose over 6 to 8 weeks, predicted future CHD events even with HbA1C levels in the normal range.17 The direct effect of glucose on ABCA1 gene expression has been examined in vitro, although results have been conflicting. Cultured HepG2 cells were grown in medium with 5.5, 12.5, or 25 mmol/L glucose. Increasing glucose concentration in the medium was associated with increasing ABCA1 mRNA levels.18 However, incubation of cultured HepG2 cells with glucose has also been shown to have no effect on ABCA1 mRNA.16 We have examined expression of the ABCA1 gene in human leukocytes obtained from healthy, nondiabetic men and inves-
From the Medical Research Council Clinical Sciences Centre, Faculty of Medicine, Imperial College, Hammersmith Hospital; and the Department of Endocrinology and Metabolic Medicine, Imperial College School of Medicine, St Mary’s Hospital, London, UK. Submitted December 23, 2002; accepted July 14, 2003. Supported by the Swiss National Science Foundation, Fellowship No. 823A-056551 for C. Albrecht and the Medical Research Council, UK. Address reprint requests to Jonathan Valabhji, MD, Department Endocrinology and Metabolic Medicine, Imperial College School of Medicine, St Mary’s Hospital, Praed Street, London W2 1NY, UK. © 2004 Elsevier Inc. All rights reserved. 0026-0495/04/5301-0032$30.00/0 doi:10.1016/j.metabol.2003.07.009 17
18
ALBRECHT ET AL
tigated its relationship to circulating glucose concentrations and to lipid, lipoprotein, and apolipoprotein parameters, including HDL-cholesterol. We hypothesized that leukocyte ABCA1 gene expression would be associated with indices of glycaemia. Peroxisome proliferator-activated receptors (PPARs) are nuclear receptors that regulate lipid and glucose metabolism and cellular differentiation. PPAR␥ activators induce ABCA1 gene expression.19 Liver X receptors (LXR) also regulate ABCA1 expression by binding to response elements in the ABCA1 gene.20 In order to investigate potential underlying mechanisms for an association between glucose and ABCA1, we also measured leukocyte mRNA expression of PPAR␥ and LXR␣. MATERIALS AND METHODS
Subjects Thirty-two healthy male subjects between 20 and 50 years of age were recruited from hospital staff. All subjects were nonsmokers, and were not known to have impaired glucose tolerance, diabetes, hyperlipidemia, or impaired renal or thyroid function. None were known to be hypertensive or to have any significant medical disorders. Subjects were assessed clinically by the same investigator, following an 8-hour overnight fast. Blood pressure was measured once using an Omron 711 automatic sphigmomanometer (Omron Healthcare, Henfield, UK) after the subject was rested sitting for 10 minutes. Venous blood samples were obtained in the overnight fasting state. Ethical committee approval was obtained from the Local Research Ethics Committee, St Mary’s NHS Trust, and informed consent was obtained from all participating subjects.
Laboratory Measurements Creatinine, HbA1c, and fasting glucose concentrations were measured using standard laboratory techniques. Cholesterol, triglyceride, and HDL-cholesterol concentrations were measured using the appropriate Olympus System Reagents (6116 and 6216, 6133, and OSR 6156, respectively; Olympus Diagnostic Systems, Southall, UK), on an Olympus AU640 analyzer system. Apolipoprotein concentrations were measured using rate immunonephelometry on an Array 360 Protein System (Beckman Coulter UK, High Wycombe, UK).
ABCA1 Gene Expression cDNA preparation. Leukocytes were isolated from 10 mL whole blood according to the Qiagen RNeasy Midi manual (Qiagen, Crawley, UK). Briefly, erythrocytes were selectively lysed in erythrocyte lysis buffer (EL; Qiagen). Leukocytes were pelleted by centrifugation at 400 ⫻ g for 10 minutes at 4°C, washed twice with EL, and resuspended in 4 ml RLT buffer (Qiagen) containing -mercaptoethanol. Samples were homogenized for 1 minute using a Polytron homogenizer (Kinematica AG, Lucerne, Switzerland) and total RNA was extracted using RNeasy midi columns (Qiagen). Contaminating DNA was removed from RNA by incubation with DNAse I (Promega, Southampton, UK). For cDNA synthesis, RNA (1 g) was transcribed with a first strand cDNA synthesis kit for reverse-transcription polymerase chain reaction (RT-PCR; Roche, Hertfordshire, UK), according to the supplier’s instructions. The RT reaction consisted of 0.08 U random hexamer primers, 20U Avium myeloblastosis virus (AMV) RT, 5 mmol/L MgCl2, 1 mmol/L deoxynucleotide mix, 40 U RNAse inhibitor, and 1x reaction buffer (Roche). For Taqman analysis, the cDNA was diluted 5-fold with nuclease-free water. Real-time quantitative RT-PCR (TaqMan). Primers and probe for TaqMan analysis of ABCA1 mRNA were designed to span 2 adjacent exons with PrimerExpress software (PE Applied Biosystems, Warrington, UK) and the reaction optimized according to PE User bulletin
no. 2 (PE Applied Biosystems, www.pebio.com). The forward primer was GGGAGGCTCCCGGAGTT in exon 3, the reverse primer was GTATAAAAGAAGCCTCCGAGCATC in exon 4, and the FAMlabeled probe, spanning exons 3 and 4, was AACTTTAACAAATCCATTGTGGCTCGCCTGT. Primers and FAM-labeled fluorescent probes for LXR alpha (5⬘-3⬘ CAAGTGTTTGCACTGCGTCT, 5⬘-3⬘ CAGGAATGTTTGCCCTTCTC, probe 5⬘-3⬘ CACTTCTAGGAGGCAGCCAC) and PPAR␥ (5⬘-3⬘ CCAGTGGTTGCAGATTACAAGTCTG, 5⬘-3⬘ TTGTAGAGCTGAGTCTTCTCAGAATAATAAG, probe 5⬘-3⬘ ACTTCAAGAGTACCAAAGTGCAATCAAAGTGGAG) were kindly provided by Drs Patel and Smith (Glaxo Smith Kline, Uxbridge, UK). Single-tube TaqMan analysis was performed on an ABI prism 7700 sequence detection system with 300 nmol/L of forward and reverse primers in the presence of 200 nmol/L 5⬘FAM-3⬘TAMRA-tagged probe for ABCA1. TaqMan measurements for LXR␣ and PPAR␥ were performed with 900 nmol/L of forward and reverse primers and 300 nmol/L 5⬘FAM-3⬘TAMRA-tagged probe. The internal standard was glyceraldehyde-3-phosphate dehydrogenase (G3PDH) mRNA, assayed with commercially supplied reagents (PE Applied Biosystems). Reactions were performed in triplicate and contained 5 L of diluted cDNA in a total volume of 25 L. Quantitation. The amount of ABCA1, LXR␣, and PPAR␥ mRNA in cells was calculated according to the relative standard curve method described in the PE User bulletin no. 2. Briefly, standard curves were determined for all target genes (ABCA1, LXR␣, PPAR␥) and G3PDH using human liver RNA samples. Target quantity was calculated from the standard curve, normalized to G3PDH (as is conventional with measurements of this kind) and expressed relative to a calibrator, which in our experimental design was defined as the mean of all ABCA1 measurements after normalization. Thus, all quantities are unitless and expressed as n-fold differences relative to the calibrator. Validation. Following extraction of mRNA, subject samples were batched. RT reactions were performed and ABCA1 gene expression was then determined for all the samples in one assay. Both RT reactions and TaqMan measurements of ABCA1 gene expression were repeated for all samples in a second assay measured on a different day. It was therefore possible to calculate the interassay coefficient of variation for measurement of ABCA1 gene expression. For subsequent analyses, the means of these 2 values for ABCA1 gene expression were used. Four subjects had very high measured levels of ABCA1 gene expression after normalization with G3PDH due to a very low signal for G3PDH in the TaqMan determinations (between 16- and 23-fold lower compared to the mean G3PDH value for the other subjects). It was important to determine whether these 4 values were true values for ABCA1 gene expression, or whether they represented either different levels of expression of the internal standard, G3PDH, or interference with the G3PDH fluorescence-based assay. A third TaqMan assay was therefore performed using a probe and oligos for -actin (PE Applied Biosystems) as internal standard. When -actin was used as the internal standard for the TaqMan determinations and analysis, the levels of ABCA1 gene expression for these 4 subjects fell within the range of values for the other subjects. It was therefore assumed that the high measured levels of ABCA1 gene expression in these 4 subjects resulted from different levels of expression of the internal standard G3PDH or an interference with the fluorescence signal. These 4 subjects were therefore excluded from the main analyses. The interassay coefficient of variation for measurement of ABCA1 gene expression was 24%. Similarily, LXR␣ and PPAR␥ expression levels in leukocytes were determined both with GAPDH and -actin as housekeeping genes and analyzed as mentioned above for ABCA1. The interassay coefficient of variation for LXR␣ and PPAR␥ gene expression were 29% and 23%, respectively.
ABCA1 EXPRESSION AND FASTING GLUCOSE
19 Table 1. Subject Characteristics (N ⴝ 32)
Parameter
Mean (or median)
Standard Deviation (or interquartile range)
Range
Age (yr) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Fasting glucose (mmol/l) HbA1c (%) Creatinine (mol/L) Total cholesterol (mmol/L) Triglyceride (mmol/l) HDL-cholesterol (mmol/L) Apolipoprotein B100 (mg/dL) Apolipoprotein A1 (mg/dL)
30 (median) 25.6 121 82 4.9 4.8 105 4.94 0.89 (median) 1.32 99 129
27-34 (interquartile range) 3.1 10 8 0.4 0.4 8 0.94 0.69-1.53 (interquartile range) 0.29 28 15
23-46 20.3-33.3 90-143 60-97 4.2-5.6 4.1-5.7 93-128 2.86-7.54 0.41-4.27 0.88-2.02 47-176 104-156
Statistical Analysis
Thirty-two subjects were recruited. Four subjects were excluded (as described in above), so that 28 subjects were included in the main analyses. Table 1 shows subject characteristics, including blood pressure, lipid, and glucose parameters. The characteristics of the 4 excluded subjects were not significantly different (data not shown). There was a strong inverse association between ABCA1 gene expression (loge-transformed) and fasting glucose concentration (r ⫽ ⫺0.49, P ⫽ .008; Fig 1). There was no significant association with HbA1c (r ⫽ ⫺0.23, P ⫽ .238). There was a significant inverse association between ABCA1 gene expression and age (r ⫽ ⫺0.39, P ⫽ .043; Fig 2). There were no signif-
icant associations between ABCA1 gene expression and HDLcholesterol concentration (r ⫽ 0.02, P ⫽ .904), apolipoprotein A1 concentration (r ⫽ ⫺0.17, P ⫽ .400), total cholesterol concentration (r ⫽ ⫺0.11, P ⫽ .587), triglyceride concentration (loge-transformed) (r ⫽ 0.06, P ⫽ .773), apolipoprotein B100 concentration (r ⫽ ⫺0.09, P ⫽ .646), systolic blood pressure (r ⫽ 0.10, P ⫽ .624), or diastolic blood pressure (r ⫽ ⫺0.07, P ⫽ .720). There was an association between fasting glucose concentration and age (r ⫽ 0.39, P ⫽ .040). In a multiple linear regression model with ABCA1 gene expression (loge-transformed) as the dependent variable and age and fasting glucose as predictors, fasting glucose remained a significant independent predictor (P ⫽ .037), whereas age did not (P ⫽ .226). There were no significant associations between fasting glucose concentration and PPAR␥ gene expression (r ⫽ 0.32, P ⫽ .113) or between fasting glucose concentration and LXR␣ gene expression (loge-transformed) (r ⫽ 0.12, P ⫽ .546). There were no significant associations between ABCA1 gene expression (loge-transformed) and PPAR␥ gene expression (r ⫽ 0.148, P ⫽ .471) or between ABCA1 gene expression (loge-transformed) and LXR␣ gene expression (r ⫽ ⫺0.34, P ⫽ .082). To confirm the validity of the association between ABCA1 gene expression and fasting glucose concentration, the association between ABCA1 gene expression measured relative to beta actin as the internal standard and fasting glucose concen-
Fig 1. ABCA1 gene expression (loge-transformed) v fasting glucose concentration (N ⴝ 28).
Fig 2. ABCA1 gene expression (loge-transformed) v age (N ⴝ 28).
Continuous variables with normal distributions are expressed as means with standard deviations. Continuous variables with skewed distributions are expressed as medians with interquartile ranges. Linear regression was used to assess the relationship between continuous variables. Residual analyses were performed: a scatterplot of the residuals versus the fitted values established homoscedasticity; the residuals were tested for normality using the Shapiro-Wilk W test. If the residuals did not have constant variance or were not normally distributed, then a variable was loge-transformed prior to regression. A 2-tailed P value less than .05 was considered significant. The Arcus Quickstat Biomedical package was used for the analyses (Longman Software Publishing, Cambridge, UK). RESULTS
20
ALBRECHT ET AL
Fig 3. ABCA1 gene expression (loge-transformed) measured using -actin as the internal standard v fasting glucose concentration (N ⴝ 32).
tration was assessed in all 32 subjects. The association remained significant (r ⫽ ⫺0.35, P ⫽ .048; Fig 3). DISCUSSION
This is the first demonstration of an association between ABCA1 gene expression and fasting glucose concentration in vivo. The fact that the significant association was with fasting glucose concentration, and not HbA1c, may reflect the fact that the time-scale for changes in ABCA1 gene expression more closely resembles that for glucose (minutes) than that for HbA1c (weeks). Results from the Norfolk cohort of the European Prospective Investigation into Cancer and Nutrition demonstrated that an index of glycemia (HbA1c), even within the normal range, predicted future CHD events.17 While an association between ABCA1 gene expression and fasting glucose concentration could link indices of glycemia and CHD events, no marker for CHD was assessed in the current study. It has been demonstrated that unsaturated fatty acids decrease ABCA1 activity in cultured macrophages and, since type 2 diabetes is associated with increased circulating nonesterified fatty acids, it has been suggested that this could explain the excess CHD risk in this condition.21 However, in the same study, the fatty acid effect was through increased ABCA1 protein degradation, and levels of ABCA1 mRNA were not affected.21 Others, however, have been able to demonstrate that unsaturated fatty acids downregulate ABCA1 mRNA.16 Unfortunately, fatty acids were not measured in the current study. Hyperglycemia is common to both type 1 and type 2 diabetes. Type 1 diabetes is associated with normal or even increased HDL-cholesterol concentrations,22 whereas type 2 diabetes is associated with decreased HDL concentrations.23 The results suggest that, if low ABCA1 gene expression is involved in the higher CHD risk of both type 1 and type 2 diabetes, this could be through an association with glucose concentration, rather than HDL-cholesterol. Despite the prespecified young age of the cohort of subjects in the present study, there was a significant univariate association between ABCA1 gene expression and age. This is the first study to demonstrate such a relationship. It has been observed that in individuals heterozygous for ABCA1 mutations, HDL-
cholesterol is inversely associated with age.24 This was assumed to be mediated by decreased ABCA1 activity, although no ABCA1 data were reported.24 The association between age and CHD is so strong that any important pathophysiological step in the disease process would also have to show a strong association with age. Age is a surrogate marker for physiological or pathological characteristics that change with time. An important challenge is to identify which physiological or pathological characteristics mediate the association between age and CHD. It is possible that decreased ABCA1 gene expression is one such characteristic and may result from increasing glucose concentrations (even within the normal range) with ageing. This is suggested by the observation in the current study that age is no longer a significant independent predictor of ABCA1 gene expression when fasting glucose concentration is considered. Although ABCA1 activity appears to play an important role in reverse cholesterol transport, it is one of many factors that contribute to the amount of cholesterol carried on HDL particles. Other important factors include lecithin:cholesterol acyltransferase and cholesterol ester transfer protein activities, lipoprotein and hepatic lipase activities, triglyceride concentrations, and activity of the hepatic scavenger receptor B-1, as well as concentrations of the apolipoproteins A1 and A2. In a murine model, discordance was seen between relative ABCA1 mRNA and protein expression in different tissues, suggesting the possibility of post-transcriptional regulation of ABCA1 expression.25 In turn ABCA1 protein levels may not be an accurate determinant of ABCA1 activity, if activity is regulated. These factors may all contribute to the lack of a demonstrable association between ABCA1 gene expression and HDL-cholesterol concentration. PPAR␥ activators induce ABCA1 gene expression and also increase apolipoprotein AI–induced cholesterol efflux from macrophages.19 PPAR␥ activation in vivo, using thiazolidinedione agents, results in reduced glucose concentrations.26 However, there was no association between glucose and levels of PPAR␥ gene expression. There was also no association between glucose and levels of LXR␣ gene expression. While this may suggest that the association between glucose and ABCA1 expression is mediated via other regulatory pathways, it should be noted that we were also unable to demonstrate associations between ABCA1 expression and expressions of PPAR␥ and LXR␣, despite their acknowledged roles in the regulation of ABCA1 expression. Given the small numbers in the study, the lack of associations between glucose and PPAR␥ expression and LXR expression may therefore represent type 2 errors. In summary, we have demonstrated an inverse association between ABCA1 gene expression and fasting glucose concentration in vivo. The association could have implications regarding the higher CHD risk associated with hyperglycemic states. ACKNOWLEDGMENT We thank Professor Nishi Chaturvedi for providing statistical advice and Caroline Chan for her skillful work in the laboratory. The authors are grateful to Dr Lisa Patel and Dr Steve Smith, Glaxo-Smith-Kline, for providing us with primers and fluorscent probes for LXR␣ and PPAR␥.
ABCA1 EXPRESSION AND FASTING GLUCOSE
21
REFERENCES 1. Miller GJ, Miller NE: Plasma high-density-lipoprotein concentration and development of ischaemic heart-disease. Lancet 1:16-19, 1975 2. Gordon DJ, Probstfield JL, Garrison RJ, et al: High-density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies. Circulation 79:8-15, 1989 3. Wang N, Silver DL, Costet P, et al: Specific binding of ApoA-I, enhanced cholesterol efflux, and altered plasma membrane morphology in cells expressing ABC1. J Biol Chem 275:33053-33058, 2000 4. Bortnick AE, Rothblat GH, Stoudt G, et al: The correlation of ATP-binding cassette 1 mRNA levels with cholesterol efflux from various cell lines. J Biol Chem 275:28634-28640, 2000 5. Oram JF, Lawn RM, Garvin MR, et al: ABCA1 is the cAMPinducible apolipoprotein receptor that mediates cholesterol secretion from macrophages. J Biol Chem 275:34508-34511, 2000 6. Brooks Wilson A, Marcil M, Clee SM, et al: Mutations in ABC1 in Tangier disease and familial high-density lipoprotein deficiency. Nat Genet 22:336-345, 1999 7. Bodzioch M, Orso E, Klucken J, et al: The gene encoding ATP-binding cassette transporter 1 is mutated in Tangier disease. Nat Genet 22:347-351, 1999 8. Rust S, Rosier M, Funke H, et al: Tangier disease is caused by mutations in the gene encoding ATP-binding cassette transporter 1. Nat Genet 22:352-355, 1999 9. Clee SM, Zwinderman AH, Engert JC, et al: Common genetic variation in ABCA1 is associated with altered lipoprotein levels and a modified risk for coronary artery disease. Circulation 103:1198-1205, 2001 10. Brousseau ME, Bodzioch M, Schaefer EJ, et al: Common variants in the gene encoding ATP-binding cassette transporter 1 in men with low HDL cholesterol levels and coronary heart disease. Atherosclerosis 154:607-611, 2001 11. Zwarts K, Clee S, Zwinderman A, et al: ABCA1 regulatory variants influence coronary artery disease independent of effects on plasma lipid levels. Clin Genet 61:115-125, 2002 12. Van Eck M, Bos I, Kaminski W, et al: Leukocyte ABCA1 controls susceptibility to atherosclerosis and macrophage recruitment into tissues. Proc Natl Acad Sci USA 99:6298-6303, 2002 13. Haghpassand M, Bourassa PA, Francone OL, et al: Monocyte/ macrophage expression of ABCA1 has minimal contribution to plasma HDL levels. J Clin Invest 108:1315-1320, 2001
14. Laing S, Swerdlow A, Slater S, et al: The British Diabetic Association Cohort Study, II: Cause-specific mortality in patients with insulin-treated diabetes mellitus. Diabet Med 16:466-471, 1999 15. Kannel WB, McGee DL: Diabetes and cardiovascular risk factors: The Framingham study. Circulation 59:8-13, 1979 16. Uehara Y, Engel T, Li Z, et al: Polyunsaturated fatty acids and acetoacetate downregulate the expression of the ATP-binding cassette transporter. Diabetes 51:2922-2928, 2002 17. Khaw KT, Wareham N, Luben R, et al: Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk). BMJ 322: 15-18, 2001 18. Tu A-Y, Albers J: Glucose regulates the transcription of human genes relevant to HDL metabolism: responsive elements for peroxisome proliferator-activated receptor are involved in the regulation of phospholipid transfer protein. Diabetes 50:1851-1856, 2001 19. Chinetti G, Lestavel S, Bocher V, et al: PPAR-alpha and PPARgamma activators induce cholesterol removal from human macrophage foam cells through stimulation of the ABCA1 pathway. Nat Med 7:53-58, 2001 20. Schwartz K, Lawn R, Wade D: ABC1 gene expression and apoA-I-mediated cholesterol efflux are regulated by LXR. Biochem Biophys Res Commun 274:794-802, 2000 21. Wang Y, Oram J: Unsaturated fatty acids inhibit cholesterol efflux from macrophages by increasing degradation of ATP-binding cassette transporter A1. J Biol Chem 277:5692-5697, 2002 22. Valabhji J, McColl A, Schachter M, et al: High-density lipoprotein composition and paraoxonase activity in type 1 diabetes. Clin Sci (Colch) 101:659-670, 2001 23. Ginsberg HN: Lipoprotein physiology in nondiabetic and diabetic states. Relationship to atherogenesis. Diabetes Care 14:839-855, 1991 24. Clee SM, Kastelein JJ, van Dam M, et al: Age and residual cholesterol efflux affect HDL cholesterol levels and coronary artery disease in ABCA1 heterozygotes. J Clin Invest 106:1263-1270, 2000 25. Wellington C, Walker E, Suarez A, et al: ABCA1 mRNA and protein distribution patterns predict multiple different roles and levels of regulation. Lab Invest 82:273-283, 2002 26. Wolffenbuttel BH, Gomis R, Squatrito S, et al: Addition of low-dose rosiglitazone to sulphonylurea therapy improves glycaemic control in type 2 diabetic patients. Diabet Med 17:40-47, 2000
A
LOW CONCENTRATION of high-density lipoprotein (HDL)-cholesterol is a powerful and independent predictor of coronary heart disease (CHD),1,2 and this is thought, though not proven, to relate to the role of HDL in reverse cholesterol transport. Recent studies have demonstrated that adenosine triphosphate (ATP)-binding cassette transporter A1 (ABCA1) is highly regulated in macrophages and directly or indirectly mediates the efflux of cholesterol and phospholipids to apolipoprotein A1, a process necessary for HDL formation, and the first step in reverse cholesterol transport.3-5 Whether ABCA1 is itself the cholesterol transporter or whether it regulates the transport process is not yet known. The identification of mutations in ABCA1 in patients with Tangier disease and familial HDL deficiency6-8 demonstrated that inadequate transport of phospholipid and cholesterol to the extracellular space results in the hypercatabolism of lipid-poor nascent HDL particles, low circulating HDL-cholesterol, and the predisposition to CHD. More common variants in the gene encoding ABCA1, such as the R219K variant, which has a carrier frequency of 46% in Europeans, are also associated with altered lipid levels and a modified risk of CHD.9 The association between ABCA1 and CHD is only partly explained by HDL-cholesterol concentrations. Indeed, in apparently normal human populations, an association between ABCA1 and HDL-cholesterol has not been clearly demonstrated. Common variants in the gene encoding ABCA1 are not associated with changes in HDL-cholesterol, despite modifying CHD risk.10 ABCA1 regulatory variants influence CHD-independent of effects on HDL-cholesterol.11 In low-density lipoprotein (LDL)-receptor knockout mice, leukocyte ABCA1 plays a critical role in the protection against atherosclerosis, but disruption of leukocyte ABCA1 function does not affect plasma HDL-cholesterol concentrations despite increasing the atherosclerotic risk.12 Experiments involving bone marrow transplantation in ABCA1-deficient and wild-type mice suggest that macrophage expression of ABCA1 makes minimal contribution to plasma HDL concentrations.13 Diabetes is associated with increased CHD risk.14,15 The risk Metabolism, Vol 53, No 1 (January), 2004: pp 17-21
is in excess of that predicted by lipid parameters, blood pressure, smoking status, and age. Whether hyperglycemia itself mediates the excess risk is unproven. ABCA1 gene expression is decreased in the liver and peritoneal macrophages of diabetic compared to control mice.16 Reduced ABCA1 function could therefore contribute to the excess CHD risk in human diabetes, and this could be independent of HDL-cholesterol. The potential importance of glucose itself is suggested by a recent prospective population-based study in men, the Norfolk cohort of the European Prospective Investigation into Cancer and Nutrition. This study has demonstrated that glycated hemoglobin (HbA1C), a measure of average circulating glucose over 6 to 8 weeks, predicted future CHD events even with HbA1C levels in the normal range.17 The direct effect of glucose on ABCA1 gene expression has been examined in vitro, although results have been conflicting. Cultured HepG2 cells were grown in medium with 5.5, 12.5, or 25 mmol/L glucose. Increasing glucose concentration in the medium was associated with increasing ABCA1 mRNA levels.18 However, incubation of cultured HepG2 cells with glucose has also been shown to have no effect on ABCA1 mRNA.16 We have examined expression of the ABCA1 gene in human leukocytes obtained from healthy, nondiabetic men and inves-
From the Medical Research Council Clinical Sciences Centre, Faculty of Medicine, Imperial College, Hammersmith Hospital; and the Department of Endocrinology and Metabolic Medicine, Imperial College School of Medicine, St Mary’s Hospital, London, UK. Submitted December 23, 2002; accepted July 14, 2003. Supported by the Swiss National Science Foundation, Fellowship No. 823A-056551 for C. Albrecht and the Medical Research Council, UK. Address reprint requests to Jonathan Valabhji, MD, Department Endocrinology and Metabolic Medicine, Imperial College School of Medicine, St Mary’s Hospital, Praed Street, London W2 1NY, UK. © 2004 Elsevier Inc. All rights reserved. 0026-0495/04/5301-0032$30.00/0 doi:10.1016/j.metabol.2003.07.009 17
18
ALBRECHT ET AL
tigated its relationship to circulating glucose concentrations and to lipid, lipoprotein, and apolipoprotein parameters, including HDL-cholesterol. We hypothesized that leukocyte ABCA1 gene expression would be associated with indices of glycaemia. Peroxisome proliferator-activated receptors (PPARs) are nuclear receptors that regulate lipid and glucose metabolism and cellular differentiation. PPAR␥ activators induce ABCA1 gene expression.19 Liver X receptors (LXR) also regulate ABCA1 expression by binding to response elements in the ABCA1 gene.20 In order to investigate potential underlying mechanisms for an association between glucose and ABCA1, we also measured leukocyte mRNA expression of PPAR␥ and LXR␣. MATERIALS AND METHODS
Subjects Thirty-two healthy male subjects between 20 and 50 years of age were recruited from hospital staff. All subjects were nonsmokers, and were not known to have impaired glucose tolerance, diabetes, hyperlipidemia, or impaired renal or thyroid function. None were known to be hypertensive or to have any significant medical disorders. Subjects were assessed clinically by the same investigator, following an 8-hour overnight fast. Blood pressure was measured once using an Omron 711 automatic sphigmomanometer (Omron Healthcare, Henfield, UK) after the subject was rested sitting for 10 minutes. Venous blood samples were obtained in the overnight fasting state. Ethical committee approval was obtained from the Local Research Ethics Committee, St Mary’s NHS Trust, and informed consent was obtained from all participating subjects.
Laboratory Measurements Creatinine, HbA1c, and fasting glucose concentrations were measured using standard laboratory techniques. Cholesterol, triglyceride, and HDL-cholesterol concentrations were measured using the appropriate Olympus System Reagents (6116 and 6216, 6133, and OSR 6156, respectively; Olympus Diagnostic Systems, Southall, UK), on an Olympus AU640 analyzer system. Apolipoprotein concentrations were measured using rate immunonephelometry on an Array 360 Protein System (Beckman Coulter UK, High Wycombe, UK).
ABCA1 Gene Expression cDNA preparation. Leukocytes were isolated from 10 mL whole blood according to the Qiagen RNeasy Midi manual (Qiagen, Crawley, UK). Briefly, erythrocytes were selectively lysed in erythrocyte lysis buffer (EL; Qiagen). Leukocytes were pelleted by centrifugation at 400 ⫻ g for 10 minutes at 4°C, washed twice with EL, and resuspended in 4 ml RLT buffer (Qiagen) containing -mercaptoethanol. Samples were homogenized for 1 minute using a Polytron homogenizer (Kinematica AG, Lucerne, Switzerland) and total RNA was extracted using RNeasy midi columns (Qiagen). Contaminating DNA was removed from RNA by incubation with DNAse I (Promega, Southampton, UK). For cDNA synthesis, RNA (1 g) was transcribed with a first strand cDNA synthesis kit for reverse-transcription polymerase chain reaction (RT-PCR; Roche, Hertfordshire, UK), according to the supplier’s instructions. The RT reaction consisted of 0.08 U random hexamer primers, 20U Avium myeloblastosis virus (AMV) RT, 5 mmol/L MgCl2, 1 mmol/L deoxynucleotide mix, 40 U RNAse inhibitor, and 1x reaction buffer (Roche). For Taqman analysis, the cDNA was diluted 5-fold with nuclease-free water. Real-time quantitative RT-PCR (TaqMan). Primers and probe for TaqMan analysis of ABCA1 mRNA were designed to span 2 adjacent exons with PrimerExpress software (PE Applied Biosystems, Warrington, UK) and the reaction optimized according to PE User bulletin
no. 2 (PE Applied Biosystems, www.pebio.com). The forward primer was GGGAGGCTCCCGGAGTT in exon 3, the reverse primer was GTATAAAAGAAGCCTCCGAGCATC in exon 4, and the FAMlabeled probe, spanning exons 3 and 4, was AACTTTAACAAATCCATTGTGGCTCGCCTGT. Primers and FAM-labeled fluorescent probes for LXR alpha (5⬘-3⬘ CAAGTGTTTGCACTGCGTCT, 5⬘-3⬘ CAGGAATGTTTGCCCTTCTC, probe 5⬘-3⬘ CACTTCTAGGAGGCAGCCAC) and PPAR␥ (5⬘-3⬘ CCAGTGGTTGCAGATTACAAGTCTG, 5⬘-3⬘ TTGTAGAGCTGAGTCTTCTCAGAATAATAAG, probe 5⬘-3⬘ ACTTCAAGAGTACCAAAGTGCAATCAAAGTGGAG) were kindly provided by Drs Patel and Smith (Glaxo Smith Kline, Uxbridge, UK). Single-tube TaqMan analysis was performed on an ABI prism 7700 sequence detection system with 300 nmol/L of forward and reverse primers in the presence of 200 nmol/L 5⬘FAM-3⬘TAMRA-tagged probe for ABCA1. TaqMan measurements for LXR␣ and PPAR␥ were performed with 900 nmol/L of forward and reverse primers and 300 nmol/L 5⬘FAM-3⬘TAMRA-tagged probe. The internal standard was glyceraldehyde-3-phosphate dehydrogenase (G3PDH) mRNA, assayed with commercially supplied reagents (PE Applied Biosystems). Reactions were performed in triplicate and contained 5 L of diluted cDNA in a total volume of 25 L. Quantitation. The amount of ABCA1, LXR␣, and PPAR␥ mRNA in cells was calculated according to the relative standard curve method described in the PE User bulletin no. 2. Briefly, standard curves were determined for all target genes (ABCA1, LXR␣, PPAR␥) and G3PDH using human liver RNA samples. Target quantity was calculated from the standard curve, normalized to G3PDH (as is conventional with measurements of this kind) and expressed relative to a calibrator, which in our experimental design was defined as the mean of all ABCA1 measurements after normalization. Thus, all quantities are unitless and expressed as n-fold differences relative to the calibrator. Validation. Following extraction of mRNA, subject samples were batched. RT reactions were performed and ABCA1 gene expression was then determined for all the samples in one assay. Both RT reactions and TaqMan measurements of ABCA1 gene expression were repeated for all samples in a second assay measured on a different day. It was therefore possible to calculate the interassay coefficient of variation for measurement of ABCA1 gene expression. For subsequent analyses, the means of these 2 values for ABCA1 gene expression were used. Four subjects had very high measured levels of ABCA1 gene expression after normalization with G3PDH due to a very low signal for G3PDH in the TaqMan determinations (between 16- and 23-fold lower compared to the mean G3PDH value for the other subjects). It was important to determine whether these 4 values were true values for ABCA1 gene expression, or whether they represented either different levels of expression of the internal standard, G3PDH, or interference with the G3PDH fluorescence-based assay. A third TaqMan assay was therefore performed using a probe and oligos for -actin (PE Applied Biosystems) as internal standard. When -actin was used as the internal standard for the TaqMan determinations and analysis, the levels of ABCA1 gene expression for these 4 subjects fell within the range of values for the other subjects. It was therefore assumed that the high measured levels of ABCA1 gene expression in these 4 subjects resulted from different levels of expression of the internal standard G3PDH or an interference with the fluorescence signal. These 4 subjects were therefore excluded from the main analyses. The interassay coefficient of variation for measurement of ABCA1 gene expression was 24%. Similarily, LXR␣ and PPAR␥ expression levels in leukocytes were determined both with GAPDH and -actin as housekeeping genes and analyzed as mentioned above for ABCA1. The interassay coefficient of variation for LXR␣ and PPAR␥ gene expression were 29% and 23%, respectively.
ABCA1 EXPRESSION AND FASTING GLUCOSE
19 Table 1. Subject Characteristics (N ⴝ 32)
Parameter
Mean (or median)
Standard Deviation (or interquartile range)
Range
Age (yr) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Fasting glucose (mmol/l) HbA1c (%) Creatinine (mol/L) Total cholesterol (mmol/L) Triglyceride (mmol/l) HDL-cholesterol (mmol/L) Apolipoprotein B100 (mg/dL) Apolipoprotein A1 (mg/dL)
30 (median) 25.6 121 82 4.9 4.8 105 4.94 0.89 (median) 1.32 99 129
27-34 (interquartile range) 3.1 10 8 0.4 0.4 8 0.94 0.69-1.53 (interquartile range) 0.29 28 15
23-46 20.3-33.3 90-143 60-97 4.2-5.6 4.1-5.7 93-128 2.86-7.54 0.41-4.27 0.88-2.02 47-176 104-156
Statistical Analysis
Thirty-two subjects were recruited. Four subjects were excluded (as described in above), so that 28 subjects were included in the main analyses. Table 1 shows subject characteristics, including blood pressure, lipid, and glucose parameters. The characteristics of the 4 excluded subjects were not significantly different (data not shown). There was a strong inverse association between ABCA1 gene expression (loge-transformed) and fasting glucose concentration (r ⫽ ⫺0.49, P ⫽ .008; Fig 1). There was no significant association with HbA1c (r ⫽ ⫺0.23, P ⫽ .238). There was a significant inverse association between ABCA1 gene expression and age (r ⫽ ⫺0.39, P ⫽ .043; Fig 2). There were no signif-
icant associations between ABCA1 gene expression and HDLcholesterol concentration (r ⫽ 0.02, P ⫽ .904), apolipoprotein A1 concentration (r ⫽ ⫺0.17, P ⫽ .400), total cholesterol concentration (r ⫽ ⫺0.11, P ⫽ .587), triglyceride concentration (loge-transformed) (r ⫽ 0.06, P ⫽ .773), apolipoprotein B100 concentration (r ⫽ ⫺0.09, P ⫽ .646), systolic blood pressure (r ⫽ 0.10, P ⫽ .624), or diastolic blood pressure (r ⫽ ⫺0.07, P ⫽ .720). There was an association between fasting glucose concentration and age (r ⫽ 0.39, P ⫽ .040). In a multiple linear regression model with ABCA1 gene expression (loge-transformed) as the dependent variable and age and fasting glucose as predictors, fasting glucose remained a significant independent predictor (P ⫽ .037), whereas age did not (P ⫽ .226). There were no significant associations between fasting glucose concentration and PPAR␥ gene expression (r ⫽ 0.32, P ⫽ .113) or between fasting glucose concentration and LXR␣ gene expression (loge-transformed) (r ⫽ 0.12, P ⫽ .546). There were no significant associations between ABCA1 gene expression (loge-transformed) and PPAR␥ gene expression (r ⫽ 0.148, P ⫽ .471) or between ABCA1 gene expression (loge-transformed) and LXR␣ gene expression (r ⫽ ⫺0.34, P ⫽ .082). To confirm the validity of the association between ABCA1 gene expression and fasting glucose concentration, the association between ABCA1 gene expression measured relative to beta actin as the internal standard and fasting glucose concen-
Fig 1. ABCA1 gene expression (loge-transformed) v fasting glucose concentration (N ⴝ 28).
Fig 2. ABCA1 gene expression (loge-transformed) v age (N ⴝ 28).
Continuous variables with normal distributions are expressed as means with standard deviations. Continuous variables with skewed distributions are expressed as medians with interquartile ranges. Linear regression was used to assess the relationship between continuous variables. Residual analyses were performed: a scatterplot of the residuals versus the fitted values established homoscedasticity; the residuals were tested for normality using the Shapiro-Wilk W test. If the residuals did not have constant variance or were not normally distributed, then a variable was loge-transformed prior to regression. A 2-tailed P value less than .05 was considered significant. The Arcus Quickstat Biomedical package was used for the analyses (Longman Software Publishing, Cambridge, UK). RESULTS
20
ALBRECHT ET AL
Fig 3. ABCA1 gene expression (loge-transformed) measured using -actin as the internal standard v fasting glucose concentration (N ⴝ 32).
tration was assessed in all 32 subjects. The association remained significant (r ⫽ ⫺0.35, P ⫽ .048; Fig 3). DISCUSSION
This is the first demonstration of an association between ABCA1 gene expression and fasting glucose concentration in vivo. The fact that the significant association was with fasting glucose concentration, and not HbA1c, may reflect the fact that the time-scale for changes in ABCA1 gene expression more closely resembles that for glucose (minutes) than that for HbA1c (weeks). Results from the Norfolk cohort of the European Prospective Investigation into Cancer and Nutrition demonstrated that an index of glycemia (HbA1c), even within the normal range, predicted future CHD events.17 While an association between ABCA1 gene expression and fasting glucose concentration could link indices of glycemia and CHD events, no marker for CHD was assessed in the current study. It has been demonstrated that unsaturated fatty acids decrease ABCA1 activity in cultured macrophages and, since type 2 diabetes is associated with increased circulating nonesterified fatty acids, it has been suggested that this could explain the excess CHD risk in this condition.21 However, in the same study, the fatty acid effect was through increased ABCA1 protein degradation, and levels of ABCA1 mRNA were not affected.21 Others, however, have been able to demonstrate that unsaturated fatty acids downregulate ABCA1 mRNA.16 Unfortunately, fatty acids were not measured in the current study. Hyperglycemia is common to both type 1 and type 2 diabetes. Type 1 diabetes is associated with normal or even increased HDL-cholesterol concentrations,22 whereas type 2 diabetes is associated with decreased HDL concentrations.23 The results suggest that, if low ABCA1 gene expression is involved in the higher CHD risk of both type 1 and type 2 diabetes, this could be through an association with glucose concentration, rather than HDL-cholesterol. Despite the prespecified young age of the cohort of subjects in the present study, there was a significant univariate association between ABCA1 gene expression and age. This is the first study to demonstrate such a relationship. It has been observed that in individuals heterozygous for ABCA1 mutations, HDL-
cholesterol is inversely associated with age.24 This was assumed to be mediated by decreased ABCA1 activity, although no ABCA1 data were reported.24 The association between age and CHD is so strong that any important pathophysiological step in the disease process would also have to show a strong association with age. Age is a surrogate marker for physiological or pathological characteristics that change with time. An important challenge is to identify which physiological or pathological characteristics mediate the association between age and CHD. It is possible that decreased ABCA1 gene expression is one such characteristic and may result from increasing glucose concentrations (even within the normal range) with ageing. This is suggested by the observation in the current study that age is no longer a significant independent predictor of ABCA1 gene expression when fasting glucose concentration is considered. Although ABCA1 activity appears to play an important role in reverse cholesterol transport, it is one of many factors that contribute to the amount of cholesterol carried on HDL particles. Other important factors include lecithin:cholesterol acyltransferase and cholesterol ester transfer protein activities, lipoprotein and hepatic lipase activities, triglyceride concentrations, and activity of the hepatic scavenger receptor B-1, as well as concentrations of the apolipoproteins A1 and A2. In a murine model, discordance was seen between relative ABCA1 mRNA and protein expression in different tissues, suggesting the possibility of post-transcriptional regulation of ABCA1 expression.25 In turn ABCA1 protein levels may not be an accurate determinant of ABCA1 activity, if activity is regulated. These factors may all contribute to the lack of a demonstrable association between ABCA1 gene expression and HDL-cholesterol concentration. PPAR␥ activators induce ABCA1 gene expression and also increase apolipoprotein AI–induced cholesterol efflux from macrophages.19 PPAR␥ activation in vivo, using thiazolidinedione agents, results in reduced glucose concentrations.26 However, there was no association between glucose and levels of PPAR␥ gene expression. There was also no association between glucose and levels of LXR␣ gene expression. While this may suggest that the association between glucose and ABCA1 expression is mediated via other regulatory pathways, it should be noted that we were also unable to demonstrate associations between ABCA1 expression and expressions of PPAR␥ and LXR␣, despite their acknowledged roles in the regulation of ABCA1 expression. Given the small numbers in the study, the lack of associations between glucose and PPAR␥ expression and LXR expression may therefore represent type 2 errors. In summary, we have demonstrated an inverse association between ABCA1 gene expression and fasting glucose concentration in vivo. The association could have implications regarding the higher CHD risk associated with hyperglycemic states. ACKNOWLEDGMENT We thank Professor Nishi Chaturvedi for providing statistical advice and Caroline Chan for her skillful work in the laboratory. The authors are grateful to Dr Lisa Patel and Dr Steve Smith, Glaxo-Smith-Kline, for providing us with primers and fluorscent probes for LXR␣ and PPAR␥.
ABCA1 EXPRESSION AND FASTING GLUCOSE
21
REFERENCES 1. Miller GJ, Miller NE: Plasma high-density-lipoprotein concentration and development of ischaemic heart-disease. Lancet 1:16-19, 1975 2. Gordon DJ, Probstfield JL, Garrison RJ, et al: High-density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies. Circulation 79:8-15, 1989 3. Wang N, Silver DL, Costet P, et al: Specific binding of ApoA-I, enhanced cholesterol efflux, and altered plasma membrane morphology in cells expressing ABC1. J Biol Chem 275:33053-33058, 2000 4. Bortnick AE, Rothblat GH, Stoudt G, et al: The correlation of ATP-binding cassette 1 mRNA levels with cholesterol efflux from various cell lines. J Biol Chem 275:28634-28640, 2000 5. Oram JF, Lawn RM, Garvin MR, et al: ABCA1 is the cAMPinducible apolipoprotein receptor that mediates cholesterol secretion from macrophages. J Biol Chem 275:34508-34511, 2000 6. Brooks Wilson A, Marcil M, Clee SM, et al: Mutations in ABC1 in Tangier disease and familial high-density lipoprotein deficiency. Nat Genet 22:336-345, 1999 7. Bodzioch M, Orso E, Klucken J, et al: The gene encoding ATP-binding cassette transporter 1 is mutated in Tangier disease. Nat Genet 22:347-351, 1999 8. Rust S, Rosier M, Funke H, et al: Tangier disease is caused by mutations in the gene encoding ATP-binding cassette transporter 1. Nat Genet 22:352-355, 1999 9. Clee SM, Zwinderman AH, Engert JC, et al: Common genetic variation in ABCA1 is associated with altered lipoprotein levels and a modified risk for coronary artery disease. Circulation 103:1198-1205, 2001 10. Brousseau ME, Bodzioch M, Schaefer EJ, et al: Common variants in the gene encoding ATP-binding cassette transporter 1 in men with low HDL cholesterol levels and coronary heart disease. Atherosclerosis 154:607-611, 2001 11. Zwarts K, Clee S, Zwinderman A, et al: ABCA1 regulatory variants influence coronary artery disease independent of effects on plasma lipid levels. Clin Genet 61:115-125, 2002 12. Van Eck M, Bos I, Kaminski W, et al: Leukocyte ABCA1 controls susceptibility to atherosclerosis and macrophage recruitment into tissues. Proc Natl Acad Sci USA 99:6298-6303, 2002 13. Haghpassand M, Bourassa PA, Francone OL, et al: Monocyte/ macrophage expression of ABCA1 has minimal contribution to plasma HDL levels. J Clin Invest 108:1315-1320, 2001
14. Laing S, Swerdlow A, Slater S, et al: The British Diabetic Association Cohort Study, II: Cause-specific mortality in patients with insulin-treated diabetes mellitus. Diabet Med 16:466-471, 1999 15. Kannel WB, McGee DL: Diabetes and cardiovascular risk factors: The Framingham study. Circulation 59:8-13, 1979 16. Uehara Y, Engel T, Li Z, et al: Polyunsaturated fatty acids and acetoacetate downregulate the expression of the ATP-binding cassette transporter. Diabetes 51:2922-2928, 2002 17. Khaw KT, Wareham N, Luben R, et al: Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of European Prospective Investigation of Cancer and Nutrition (EPIC-Norfolk). BMJ 322: 15-18, 2001 18. Tu A-Y, Albers J: Glucose regulates the transcription of human genes relevant to HDL metabolism: responsive elements for peroxisome proliferator-activated receptor are involved in the regulation of phospholipid transfer protein. Diabetes 50:1851-1856, 2001 19. Chinetti G, Lestavel S, Bocher V, et al: PPAR-alpha and PPARgamma activators induce cholesterol removal from human macrophage foam cells through stimulation of the ABCA1 pathway. Nat Med 7:53-58, 2001 20. Schwartz K, Lawn R, Wade D: ABC1 gene expression and apoA-I-mediated cholesterol efflux are regulated by LXR. Biochem Biophys Res Commun 274:794-802, 2000 21. Wang Y, Oram J: Unsaturated fatty acids inhibit cholesterol efflux from macrophages by increasing degradation of ATP-binding cassette transporter A1. J Biol Chem 277:5692-5697, 2002 22. Valabhji J, McColl A, Schachter M, et al: High-density lipoprotein composition and paraoxonase activity in type 1 diabetes. Clin Sci (Colch) 101:659-670, 2001 23. Ginsberg HN: Lipoprotein physiology in nondiabetic and diabetic states. Relationship to atherogenesis. Diabetes Care 14:839-855, 1991 24. Clee SM, Kastelein JJ, van Dam M, et al: Age and residual cholesterol efflux affect HDL cholesterol levels and coronary artery disease in ABCA1 heterozygotes. J Clin Invest 106:1263-1270, 2000 25. Wellington C, Walker E, Suarez A, et al: ABCA1 mRNA and protein distribution patterns predict multiple different roles and levels of regulation. Lab Invest 82:273-283, 2002 26. Wolffenbuttel BH, Gomis R, Squatrito S, et al: Addition of low-dose rosiglitazone to sulphonylurea therapy improves glycaemic control in type 2 diabetic patients. Diabet Med 17:40-47, 2000